AI Playbook
for Manufacturing

You are a production planner optimizing the weekly schedule.

Current schedule:
[PASTE: Job # | Product | Qty | Due date | Line/machine | Setup time (hrs) | Run time (hrs) | Priority]

Constraints:
[LIST: Available hours per line, shift pattern, material shortages, changeover times between product families]

Optimize to:
1) Meet all due dates — flag any that cannot be met; state reason
2) Minimize total changeover time — show before vs. after comparison and hours saved
3) Group same-family products to reduce changeovers
4) Balance line utilization — flag lines >90% (constraint risk) or <60% (waste)
5) Flag jobs blocked by material shortages with expected days of impact

Output: Optimized schedule table + summary of hours saved. At-risk due dates listed separately with reason.

You are an operations manager assessing production capacity for the next quarter.

Demand data:
[PASTE: Product family | Forecasted units per week | Standard hours per unit | Line/machine required]

Capacity data:
[PASTE: Line/machine | Planned available hours per week | Current OEE %]

For each line:
1) Calculate effective capacity = Available hours × OEE %
2) Compare to required hours — identify over-capacity weeks (>85% utilized) and under-capacity weeks (<60%)
3) For over-capacity: options — overtime, additional shift, outsource, defer lower-priority work; estimate cost of each
4) For under-capacity: options — maintenance windows, training, cross-training, additional product runs
5) Flag the weeks and lines where demand exceeds effective capacity

Output: Capacity summary table by week and line. Constraint calendar. Recommended actions with estimated cost.

You are a production supervisor building the hour-by-hour plan for the shift.

Shift data:
[PASTE: Shift date/time | Line | Jobs to run (job # / product / qty) | Available operators | Known machine issues or changeovers]

Build the short-interval schedule:
1) Assign jobs to 1–2 hour time slots based on run rates and changeover sequence
2) Set production targets per interval — units per hour based on standard rate
3) Identify the critical hour — the interval most likely to cause end-of-shift shortfall
4) Note planned downtime (changeovers, breaks, PM) and confirm remaining run time is sufficient
5) Handover note — what does the next supervisor need to know before shift start?

Output: Hour-by-hour schedule table. Target units per interval. Critical alert for supervisor.

You are a production manager reviewing daily output vs. plan.

Output data:
[PASTE: Line/machine | Product | Planned units | Actual units | Variance units | Variance % | Downtime (mins) | Scrap units]

For each line below plan:
1) Calculate efficiency % = Actual ÷ Planned × 100
2) Break down shortfall: downtime losses / speed losses / scrap losses / changeover overrun
3) Identify primary cause — be specific (machine name, fault type, product, operator issue)
4) Recovery plan — can the shortfall be recovered in the next shift? What would it require?
5) Flag any line with efficiency <80% for 2+ consecutive days — this is a trend, not a blip

Output: Daily variance report. Traffic light per line: Green ≥95% / Amber 80–94% / Red <80%. Recommended action for each Red line.

You are a manufacturing engineer analyzing production run rates against standards.

Run rate data:
[PASTE: Product | Machine/line | Standard rate (units/hr) | Actual rate (units/hr) | Period | Operator count]

For each product/line combination:
1) Performance % = Actual rate ÷ Standard rate × 100
2) Flag products running consistently below 85% of standard
3) Identify the gap cause: mechanical (machine speed reduced) / manning (fewer operators) / method (process not followed) / standard error
4) Estimate throughput lost — units and $ value (if unit margin available)
5) Recommend investigation steps for lowest-performing items

Output: Run rate analysis table. Priority list for engineering or management review.

You are an industrial engineer identifying production bottlenecks.

Process flow data:
[PASTE: Process step | Cycle time (mins/unit) | Available time per shift | Number of machines/operators | Current WIP queue at this step]

Apply Theory of Constraints analysis:
1) Identify the bottleneck — highest utilization or largest WIP queue
2) Calculate theoretical throughput rate limited by the bottleneck
3) Calculate throughput lost vs. potential if bottleneck were resolved
4) For the bottleneck: recommend exploitation options (maximize output now) and elevation options (add capacity)
5) Check for subordination issues — are non-bottleneck steps starving or flooding the bottleneck?

Output: Process utilization table. Bottleneck identified with evidence. Three specific recommendations to increase throughput.

You are a production manager clearing a work order backlog.

Backlog data:
[PASTE: WO # | Product | Qty | Customer due date | Key account? (yes/no) | Line required | Estimated hours remaining | Status]

Prioritize using:
1) Customer due date — earliest first
2) Key accounts — elevate over standard at same due date
3) Line grouping — sequence by line to minimize changeovers
4) Flag orders that will miss due date given current capacity — state estimated delay
5) Identify orders requiring expedite action today to avoid customer impact

Output: Prioritized work order sequence by line. Missed due date list with estimated delay. Expedite flags for immediate action.

You are a lean engineer analyzing changeover performance using SMED principles.

Changeover data:
[PASTE: Line | From product | To product | Total changeover time (mins) | Internal time (machine stopped) | External time (prep while running) | Date | Operator]

Apply SMED:
1) Identify activities that can convert from internal to external (prep while machine is still running)
2) Simplify and standardize remaining internal activities
3) Calculate average changeover time vs. target — gap in minutes
4) For the longest changeovers: top 3 time-consuming steps and specific improvements for each
5) Estimate total production hours recovered per week if average changeover reduced by 25%

Output: SMED analysis table. Top 5 improvement actions ranked by time savings. Monthly throughput uplift.

You are a production planner reviewing MRP exception messages.

MRP exceptions:
[PASTE: Item | Exception type | Suggested action | Quantity | Date | Current status]

For each exception type:
1) Reschedule in — demand moved earlier; action: confirm supply can move forward
2) Reschedule out — demand moved later; action: push supply order to reduce inventory
3) Cancel — supply order no longer needed; action: cancel to free capacity/material
4) New planned order — demand with no supply; confirm if auto-release is appropriate
5) Past due — supply or demand past due; assess impact and recovery plan

For each exception: Accept the MRP suggestion / Override with reason / Escalate.
Flag: any exception affecting customer-facing orders — highest priority.

Output: Exception action list — Accept / Override / Escalate — with reason for each.

You are a master scheduler preparing the monthly MPS review.

Data:
[PASTE: Product | Forecasted demand (next 3 months by week) | Confirmed orders | Current finished goods inventory | Production plan (next 3 months by week) | Safety stock target]

Review for:
1) Demand vs. supply gaps — weeks where production plan doesn't cover forecast + safety stock
2) Over-planned weeks — production exceeds demand; flag inventory build-up risk
3) Demand changes vs. last month — significant forecast changes requiring schedule adjustment
4) Frozen zone violations — changes inside the [X-week] frozen horizon disrupting confirmed schedules
5) Customer order coverage — are all confirmed orders covered?

Output: MPS review summary. Gaps and over-plans by week. Recommended adjustments. Items requiring S&OP team decision.

You are a production planning manager building the pre-production schedule for a new product.

Launch data:
[DESCRIBE: Product, target launch date, first production quantity, line/machine, tooling required, key raw materials, training requirements]

Work backward from launch date:
1) List all pre-production tasks required: tooling validation, material qualification, trial runs, training, first article inspection
2) Assign owner and duration to each task
3) Identify the critical path — tasks that if delayed will push the launch date
4) Flag long-lead-time items needing immediate action
5) Define go/no-go criteria — what must be true before first production run is approved?

Output: Launch schedule table with critical path highlighted. Immediate action list. Go/no-go checklist.

You are a production manager planning for seasonal demand peaks.

Data:
[PASTE: Month | Forecasted demand (units) | Available production days | Line capacity (units/day) | Current finished goods inventory]

Plan the seasonal buildup:
1) Identify months where demand exceeds normal production capacity
2) Calculate advance inventory build needed before peak season
3) Determine when build-ahead must start and on which products
4) Identify storage constraints — will inventory exceed warehouse capacity?
5) Assess labor implications — temp workers or overtime needed during peak?

Output: Month-by-month production plan. Build-ahead quantities. Storage peak. Hiring/overtime trigger dates.